Would be nice to have an option to normalize your libraries audio. That way you don't have to worry about a chart blowing your eardrum out, but you also won't have to raise it for quiet songs.
#Normalize Audio
alpine eagle
leaden mirage
this is from the clone hero subreddit, but i'm going to try this tomorrow after making a backup of my files- https://www.reddit.com/r/CloneHero/comments/7wrwjr/ive_made_a_program_that_normalizes_your_song/
west nimbus
Maybe we could do it on the fly, I would like this feature too
leaden mirage
on the fly would be awesome. i tested that one with a small subest of files and it limits people to need twice the space available since it copies the songs to a new folder
for me i have like 200+gb of songs so i'd need minimum 400gb space to run it
alpine eagle
this is from the clone hero subreddit, but i'm going to try this tomorrow after ...
problem is it doesn't work on bass.ogg or preview.ogg, so it's not exactly a perfect solution
floral breach
i think YARG already has a compressor but i could be wrong
flint lava
west nimbus
There is a BASS addon for measuring loudness using sophisticated methods, this would allow us to do a scan right before playing the song and adjust the volume ahead of time
https://www.un4seen.com/doc/#bassloud/BASS_Loudness_GetLevelMulti.html
However it's not supported by ManagedBass C# lib yet so that would have to be done first probably, or we just write our own internal wrapper
TL;DR it can be done without too much effort
lone sinew
it is insufficiently speedy, sadly
west nimbus
Really? I haven't tried it but I can't imagine it takes long to measure loudness of 3 seconds of audio
we could also roll our own DSP that does normalization, assuming we don't crush dynamic range too hard doing that. That might be an even better solution anyway
ffmpeg can measure LUFS in like 1 second, I think we could do the measurement right before starting the song
ffmpeg -i vocals.ogg -af ebur128 -f null -
lone sinew
it needs to be a few milliseconds if we're going to do it before starting a song
west nimbus
I see. No loading screens or anything allowed before the song plays?
lone sinew
Preferably not. Like it's fine if the first time you play a song after an update it has to recompile shaders or something and it takes a second, but in general the idea is for it to be as snappy as is reasonably possible
west nimbus
I don't wanna get too wild with it but we could run the lufs measurement while the song is playing and adjust gain when its done, the first second of the song is usually silence anyway. I guess with a 6 stem song that's 6x as long to measure everything. So yeah we'd have to think about ways to normalize quickly.
alpine eagle
Ugh I want this so bad lol